Efficient Computation Offloading with Energy Consumption Constraint for Multi-Cloud System

Abstract

With the increasing functionalities of mobile terminals, computation offloading has become a good way to alleviate the limitation of terminal computing resource and improve terminal performance. In this paper, we propose an efficient computation offloading strategy for multi-cloud system to minimize makespan with terminal energy consumption constraint. The proposed strategy first sorts tasks and establishes paths. Then tasks on each established path are counted as an integrated task and assigned to a worker node with less completion time iteratively. After the initial task assignment, performance optimization is performed to satisfy energy consumption constraint and further decrease makespan. Experimental results show that the proposed approach has better performance compared to genetic algorithm and greedy algorithm.